The overriding scientific goal of this application for Research Units on Pediatric Psychopharmacology and Psychosocial Interventions (RUPP-PI) is to provide reliable estimates of mental health outcomes in mentally ill children and adolescents treated with widely used single and combined psychosocial and drug treatments for which randomized evidence is absent. The pragmatic goal of the application is to make the case that Duke University is ideally suited to participate in RUPP-PI. To meet these linked goals, we structured the application as follows: First, we define our specific aims in the context of the goals of the RFA. Second, we summarize the literature on the treatment of mentally ill youth, pointing out gaps in the literature that justify our specific aims. Third, we document the exceptional expertise and leadership experience that Duke brings to RUPP-PI. We specifically highlight our capability in research design, trials management, psychometrics, family and individual CBT, psychopharmacology, and, most importantly, our ability to work collaboratively on multidisciplinary, multicenter comparative treatment trials. Fourth, we propose a research protocol designed to illustrate our expertise in trial design and to point toward our relatively unique experience in evidence-based family therapy. Our work in the RUPP fluvoxamine anxiety trial suggests that partial responders to SSRI trials are not uncommon. Work by us and others suggests that the addition of a family treatment component to individual CBT improves outcomes in children with internalizing disorders. In this context, we propose to use a balanced 5 (site) by 3 (treatment) by 6 (repeated measures) experimental design to answer the following question: "In children and adolescents age 7-14 with separation, generalized or social anxiety who are partial responders to an SSRI, does the addition of CBT or CBT/FAM as compared to continued treatment with the SSRI (cSSRI) improve anxiety and functional outcomes acutely and at 6 month follow-up." Fifth, we propose several mechanisms to accomplish the training aim in the RFA. Finally, we document our willingness to work with other sites in selecting trial(s) to run on RUPP-PI and, in so doing, highlight the senior leadership and communication skills that we bring to RUPP-PI, skills that are critical to insuring that network members come together as stakeholders for the chosen scientific agenda.